Twenty-six studies were included in meta-analyses, including 16 controlled clinical trials (n=2,467 patients) and 10 before-and-after studies (n=38,581 patients). Fourteen clinical controlled trials were randomised controlled trials. The quality of controlled clinical trials ranged from 1 to 3 points. The follow-up duration ranged from one month to five years.
Patient-level interventions were associated with a significant reduction in all-cause mortality at follow-up (RR 0.78, 95% CI 0.71 to 0.86; 18 studies), all-cause mortality at one-year follow-up (RR 0.79, 95% CI 0.69 to 0.92; 14 studies), and re-admission rate (RR 0.84, 95% CI 0.73 to 0.98; 10 studies) compared with no intervention (control). No significant heterogeneity was observed for these outcomes.
There was no significant difference in reinfarction rate between the intervention and control groups. Significant heterogeneity was observed for this outcome (I2=90%).
Subgroup analyses by study type showed that patient-level interventions were associated with a significant reduction in all-cause mortality at one-year follow-up when pooling before-after studies (RR 0.77, 95% CI 0.66 to 0.90; five studies), but a non-significant reduction in all-cause mortality at one-year follow-up when pooling controlled clinical studies.
Funnel plots and Begg's test did not suggest significant publication bias.
Results of other secondary outcomes and subgroup analyses were also reported.